Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.user > #248046 > unrolled thread

USB Debian 11 installer and target drive device node [was fstab problem]

Started byDavid Christensen <dpchrist@holgerdanske.com>
First post2022-05-09 20:40 +0200
Last post2022-05-10 02:40 +0200
Articles 4 — 3 participants

Back to article view | Back to linux.debian.user

Contents

  USB Debian 11 installer and target drive device node [was fstab  problem] David Christensen <dpchrist@holgerdanske.com> - 2022-05-09 20:40 +0200
    Re: USB Debian 11 installer and target drive device node [was fstab problem] "Thomas Schmitt" <scdbackup@gmx.net> - 2022-05-09 21:20 +0200
    Re: USB Debian 11 installer and target drive device node [was fstab  problem] David Wright <deblis@lionunicorn.co.uk> - 2022-05-09 22:00 +0200
      Re: USB Debian 11 installer and target drive device node [was fstab  problem] David Christensen <dpchrist@holgerdanske.com> - 2022-05-10 02:40 +0200

#248046 — USB Debian 11 installer and target drive device node [was fstab problem]

FromDavid Christensen <dpchrist@holgerdanske.com>
Date2022-05-09 20:40 +0200
SubjectUSB Debian 11 installer and target drive device node [was fstab problem]
Message-ID<Elgsy-enso-11@gated-at.bofh.it>
On 5/9/22 10:21, David Wright wrote:
 > On Sun 08 May 2022 at 23:39:31 (-0700), David Christensen wrote:

 >> As noted by another reader in another thread, the Debian 11 installer
 >> appears to always assign /dev/sda to d-i USB installation media; in
 >> spite of decades of standard practice of using the first drive node
 >> for the target drive.  Thus, the target drive may have been /dev/sdb
 >> at installation time, resulting in "sdb" in crypttab(5) and/or
 >> fstab(5) entries rather than the conventional "sda".

 > This is news to me, and I'm not sure why you blame the d-i's going
 > against decades of precedent. Yes, it can happen, and I guess it might
 > be the result of the way the buses are configured inside the machine.
 > Gone are the simple days of PATAs on hda.
 >
 > Anyway, here are six machines: three are laptops, three EFIs, one AiO,
 > one i386, all GPT, all netinst, variously 11.1–3 with firmware. All
 > except the last have the installer on sdb, and the reason there is
 > obvious.
 >
 > The sole occasion where the devices have been named the "wrong" way
 > round is when I installed Debian on an EFI laptop onto a virgin
 > external hard drive. If memory serves, the d-i was connected through
 > a hub to the USB-C port, the drive was on the single USB3 port.
 > I think.
 >
 > $ grep -e sda -e sdb -e nvme /var/log/installer/syslog | sed -e 
'/CD-ROM/{p; q}'
 > Jan  1 16:29:18 kernel: [    1.815571] sd 1:0:0:0: [sda] 976773168 
512-byte logical blocks: (500 GB/466 GiB)
 > Jan  1 16:29:18 kernel: [    1.816244] sd 1:0:0:0: [sda] 4096-byte 
physical blocks
 > Jan  1 16:29:18 kernel: [    1.817495] sd 1:0:0:0: [sda] Write 
Protect is off
 > Jan  1 16:29:18 kernel: [    1.819103] sd 1:0:0:0: [sda] Mode Sense: 
00 3a 00 00
 > Jan  1 16:29:18 kernel: [    1.819791] sd 1:0:0:0: [sda] Write cache: 
enabled, read cache: enabled, doesn't support DPO or FUA
 > Jan  1 16:29:18 kernel: [    1.853168]  sda: sda1 sda2 sda3 sda4 sda5 
sda6 sda7 sda8 sda9 sda10 sda11
 > Jan  1 16:29:18 kernel: [    1.875612] sd 1:0:0:0: [sda] Attached 
SCSI disk
 > Jan  1 16:29:18 kernel: [    3.422080] sd 3:0:0:0: [sdb] 2006528 
512-byte logical blocks: (1.03 GB/980 MiB)
 > Jan  1 16:29:18 kernel: [    3.422370] sd 3:0:0:0: [sdb] Write 
Protect is off
 > Jan  1 16:29:18 kernel: [    3.422372] sd 3:0:0:0: [sdb] Mode Sense: 
03 00 00 00
 > Jan  1 16:29:18 kernel: [    3.422673] sd 3:0:0:0: [sdb] No Caching 
mode page found
 > Jan  1 16:29:18 kernel: [    3.422674] sd 3:0:0:0: [sdb] Assuming 
drive cache: write through
 > Jan  1 16:29:18 kernel: [    3.453581]  sdb: sdb1 sdb2
 > Jan  1 16:29:18 kernel: [    3.490735] sd 3:0:0:0: [sdb] Attached 
SCSI removable disk
 > Jan  1 16:29:55 cdrom-detect: CD-ROM mount succeeded: 
device=/dev/sdb1 fstype=iso9660
 > Jan  1 16:29:55 cdrom-detect: CD-ROM mount succeeded: 
device=/dev/sdb1 fstype=iso9660
 > $
 >
 > $ grep -e sda -e sdb -e nvme /var/log/installer/syslog | sed -e 
'/CD-ROM/{p; q}'
 > Jan 25 21:12:08 kernel: [    4.310937] sd 0:0:0:0: [sda] 976773168 
512-byte logical blocks: (500 GB/466 GiB)
 > Jan 25 21:12:08 kernel: [    4.310951] sd 0:0:0:0: [sda] Write 
Protect is off
 > Jan 25 21:12:08 kernel: [    4.310952] sd 0:0:0:0: [sda] Mode Sense: 
00 3a 00 00
 > Jan 25 21:12:08 kernel: [    4.310972] sd 0:0:0:0: [sda] Write cache: 
enabled, read cache: enabled, doesn't support DPO or FUA
 > Jan 25 21:12:08 kernel: [    4.681368]  sda: sda1 sda2 sda3 sda4 sda5 
sda6
 > Jan 25 21:12:08 kernel: [    4.839485] sd 0:0:0:0: [sda] Attached 
SCSI disk
 > Jan 25 21:12:08 kernel: [    5.617893] sd 6:0:0:0: [sdb] 2006528 
512-byte logical blocks: (1.03 GB/980 MiB)
 > Jan 25 21:12:08 kernel: [    5.637613] sd 6:0:0:0: [sdb] Write 
Protect is off
 > Jan 25 21:12:08 kernel: [    5.655705] sd 6:0:0:0: [sdb] Mode Sense: 
03 00 00 00
 > Jan 25 21:12:08 kernel: [    5.656965] sd 6:0:0:0: [sdb] No Caching 
mode page found
 > Jan 25 21:12:08 kernel: [    5.674988] sd 6:0:0:0: [sdb] Assuming 
drive cache: write through
 > Jan 25 21:12:08 kernel: [    5.718770]  sdb: sdb1 sdb2
 > Jan 25 21:12:08 kernel: [    5.762284] sd 6:0:0:0: [sdb] Attached 
SCSI removable disk
 > Jan 25 21:13:04 cdrom-detect: CD-ROM mount succeeded: 
device=/dev/sdb1 fstype=iso9660
 > Jan 25 21:13:04 cdrom-detect: CD-ROM mount succeeded: 
device=/dev/sdb1 fstype=iso9660
 > $
 >
 > $ grep -e sda -e sdb -e nvme /var/log/installer/syslog | sed -e 
'/CD-ROM/{p; q}'
 > Feb  1 17:32:10 kernel: [    5.562125] sd 0:0:0:0: [sda] 976773168 
512-byte logical blocks: (500 GB/466 GiB)
 > Feb  1 17:32:10 kernel: [    5.569327] sd 0:0:0:0: [sda] Write 
Protect is off
 > Feb  1 17:32:10 kernel: [    5.576324] sd 0:0:0:0: [sda] Mode Sense: 
00 3a 00 00
 > Feb  1 17:32:10 kernel: [    5.576399] sd 0:0:0:0: [sda] Write cache: 
enabled, read cache: enabled, doesn't support DPO or FUA
 > Feb  1 17:32:10 kernel: [    5.646877]  sda: sda1 sda2 sda3 sda4 sda5 
sda6
 > Feb  1 17:32:10 kernel: [    5.667737] sd 0:0:0:0: [sda] Attached 
SCSI disk
 > Feb  1 17:32:10 kernel: [    5.834514] sd 4:0:0:0: [sdb] 2006528 
512-byte logical blocks: (1.03 GB/980 MiB)
 > Feb  1 17:32:10 kernel: [    5.849013] sd 4:0:0:0: [sdb] Write 
Protect is off
 > Feb  1 17:32:10 kernel: [    5.855644] sd 4:0:0:0: [sdb] Mode Sense: 
03 00 00 00
 > Feb  1 17:32:10 kernel: [    5.856638] sd 4:0:0:0: [sdb] No Caching 
mode page found
 > Feb  1 17:32:10 kernel: [    5.863141] sd 4:0:0:0: [sdb] Assuming 
drive cache: write through
 > Feb  1 17:32:10 kernel: [    5.889163]  sdb: sdb1 sdb2
 > Feb  1 17:32:10 kernel: [    5.914390] sd 4:0:0:0: [sdb] Attached 
SCSI removable disk
 > Feb  1 17:32:45 cdrom-detect: CD-ROM mount succeeded: 
device=/dev/sdb1 fstype=iso9660
 > Feb  1 17:32:45 cdrom-detect: CD-ROM mount succeeded: 
device=/dev/sdb1 fstype=iso9660
 > $
 >
 > $ grep -e sda -e sdb -e nvme /var/log/installer/syslog | sed -e 
'/CD-ROM/{p; q}'
 > Apr 17 01:01:15 kernel: [    1.835596] sd 0:0:0:0: [sda] 976773168 
512-byte logical blocks: (500 GB/466 GiB)
 > Apr 17 01:01:15 kernel: [    1.837196] sd 0:0:0:0: [sda] 4096-byte 
physical blocks
 > Apr 17 01:01:15 kernel: [    1.838100] sd 0:0:0:0: [sda] Write 
Protect is off
 > Apr 17 01:01:15 kernel: [    1.838967] sd 0:0:0:0: [sda] Mode Sense: 
00 3a 00 00
 > Apr 17 01:01:15 kernel: [    1.838992] sd 0:0:0:0: [sda] Write cache: 
enabled, read cache: enabled, doesn't support DPO or FUA
 > Apr 17 01:01:15 kernel: [    1.863206]  sda: sda1 sda2 sda3 sda4 sda5 
sda6
 > Apr 17 01:01:15 kernel: [    1.888613] sd 0:0:0:0: [sda] Attached 
SCSI disk
 > Apr 17 01:01:16 kernel: [    3.120312] sd 6:0:0:0: [sdb] 2006528 
512-byte logical blocks: (1.03 GB/980 MiB)
 > Apr 17 01:01:16 kernel: [    3.120785] sd 6:0:0:0: [sdb] Write 
Protect is off
 > Apr 17 01:01:16 kernel: [    3.120790] sd 6:0:0:0: [sdb] Mode Sense: 
03 00 00 00
 > Apr 17 01:01:16 kernel: [    3.121108] sd 6:0:0:0: [sdb] No Caching 
mode page found
 > Apr 17 01:01:16 kernel: [    3.121111] sd 6:0:0:0: [sdb] Assuming 
drive cache: write through
 > Apr 17 01:01:16 kernel: [    3.143895]  sdb: sdb1 sdb2 sdb3
 > Apr 17 01:01:16 kernel: [    3.157941] sd 6:0:0:0: [sdb] Attached 
SCSI removable disk
 > Apr 17 01:02:28 cdrom-detect: CD-ROM mount succeeded: 
device=/dev/sdb1 fstype=iso9660
 > Apr 17 01:02:28 cdrom-detect: CD-ROM mount succeeded: 
device=/dev/sdb1 fstype=iso9660
 > $
 >
 > $ grep -e sda -e sdb -e nvme /var/log/installer/syslog | sed -e 
'/CD-ROM/{p; q}'
 > Apr 21 18:43:48 kernel: [    4.728919] sd 0:0:0:0: [sda] 117210240 
512-byte logical blocks: (60.0 GB/55.9 GiB)
 > Apr 21 18:43:48 kernel: [    4.781174] sd 0:0:0:0: [sda] Write 
Protect is off
 > Apr 21 18:43:48 kernel: [    4.791354] sd 0:0:0:0: [sda] Mode Sense: 
00 3a 00 00
 > Apr 21 18:43:48 kernel: [    4.853619] sd 0:0:0:0: [sda] Write cache: 
enabled, read cache: enabled, doesn't support DPO or FUA
 > Apr 21 18:43:48 kernel: [    5.284602]  sda: sda1 sda2 sda3 sda4
 > Apr 21 18:43:48 kernel: [    5.320578] sd 0:0:0:0: [sda] Attached 
SCSI disk
 > Apr 21 18:43:48 kernel: [    6.301453] sd 2:0:0:0: [sdb] 2006528 
512-byte logical blocks: (1.03 GB/980 MiB)
 > Apr 21 18:43:48 kernel: [    6.314811] sd 2:0:0:0: [sdb] Write 
Protect is off
 > Apr 21 18:43:48 kernel: [    6.326835] sd 2:0:0:0: [sdb] Mode Sense: 
03 00 00 00
 > Apr 21 18:43:48 kernel: [    6.328822] sd 2:0:0:0: [sdb] No Caching 
mode page found
 > Apr 21 18:43:48 kernel: [    6.340627] sd 2:0:0:0: [sdb] Assuming 
drive cache: write through
 > Apr 21 18:43:48 kernel: [    6.363150]  sdb: sdb1 sdb2 sdb3
 > Apr 21 18:43:48 kernel: [    6.387070] sd 2:0:0:0: [sdb] Attached 
SCSI removable disk
 > Apr 21 18:44:19 cdrom-detect: CD-ROM mount succeeded: 
device=/dev/sdb1 fstype=iso9660
 > Apr 21 18:44:19 cdrom-detect: CD-ROM mount succeeded: 
device=/dev/sdb1 fstype=iso9660
 > $
 >
 > $ grep -e sda -e sdb -e nvme /var/log/installer/syslog | sed -e 
'/CD-ROM/{p; q}'
 > Dec 13 03:02:14 kernel: [    0.874287] nvme nvme0: pci function 
0000:3d:00.0
 > Dec 13 03:02:14 kernel: [    0.890575] nvme nvme0: 4/0/0 
default/read/poll queues
 > Dec 13 03:02:14 kernel: [    0.892293]  nvme0n1: p1 p2 p3 p4
 > Dec 13 03:02:15 kernel: [    3.106803] sd 0:0:0:0: [sda] 2006528 
512-byte logical blocks: (1.03 GB/980 MiB)
 > Dec 13 03:02:15 kernel: [    3.107444] sd 0:0:0:0: [sda] Write 
Protect is off
 > Dec 13 03:02:15 kernel: [    3.107450] sd 0:0:0:0: [sda] Mode Sense: 
03 00 00 00
 > Dec 13 03:02:15 kernel: [    3.107907] sd 0:0:0:0: [sda] No Caching 
mode page found
 > Dec 13 03:02:15 kernel: [    3.107913] sd 0:0:0:0: [sda] Assuming 
drive cache: write through
 > Dec 13 03:02:15 kernel: [    3.125917]  sda: sda1 sda2
 > Dec 13 03:02:16 kernel: [    3.155962] sd 0:0:0:0: [sda] Attached 
SCSI removable disk
 > Dec 13 03:02:48 cdrom-detect: CD-ROM mount succeeded: 
device=/dev/sda1 fstype=iso9660
 > Dec 13 03:02:48 cdrom-detect: CD-ROM mount succeeded: 
device=/dev/sda1 fstype=iso9660
 > $


I am unsure what is obvious about the last three.  Please clarify.


Did you install from a USB flash drive or from optical media?  If 
optical, was the drive internal or external, and what was the interface?


David

[toc] | [next] | [standalone]

#248048 — Re: USB Debian 11 installer and target drive device node [was fstab problem]

From"Thomas Schmitt" <scdbackup@gmx.net>
Date2022-05-09 21:20 +0200
SubjectRe: USB Debian 11 installer and target drive device node [was fstab problem]
Message-ID<Elh5f-enUh-3@gated-at.bofh.it>
In reply to#248046
Hi,

David Wright wrote:
> > Dec 13 03:02:16 kernel: [    3.155962] sd 0:0:0:0: [sda] Attached SCSI removable disk
> > Dec 13 03:02:48 cdrom-detect: CD-ROM mount succeeded: device=/dev/sda1 fstype=iso9660
> > Dec 13 03:02:48 cdrom-detect: CD-ROM mount succeeded: device=/dev/sda1 fstype=iso9660

David Christensen wrote:
> Did you install from a USB flash drive or from optical media?

sda is not a real optical medium and Linux would not accept a partition
table from a real optical medium.


I assume the message comes from
  https://sources.debian.org/src/cdrom-detect/1.98/debian/cdrom-detect.postinst/?hl=23#L23
or
  https://sources.debian.org/src/cdrom-detect/1.98/debian/cdrom-detect.postinst/?hl=216#L216
when it tries to find the ISO 9660 filesystem from which the running
system was booted.

(The confusing messages are excusable. Debian ISOs are originally supposed
to boot from CD media. USB sticks are a novelty from 2011. :))


Have a nice day :)

Thomas

[toc] | [prev] | [next] | [standalone]

#248050

FromDavid Wright <deblis@lionunicorn.co.uk>
Date2022-05-09 22:00 +0200
Message-ID<ElhHY-eo7a-19@gated-at.bofh.it>
In reply to#248046
On Mon 09 May 2022 at 11:30:39 (-0700), David Christensen wrote:
> On 5/9/22 10:21, David Wright wrote:
> > On Sun 08 May 2022 at 23:39:31 (-0700), David Christensen wrote:
> 
> >> As noted by another reader in another thread, the Debian 11 installer
> >> appears to always assign /dev/sda to d-i USB installation media; in
> >> spite of decades of standard practice of using the first drive node
> >> for the target drive.  Thus, the target drive may have been /dev/sdb
> >> at installation time, resulting in "sdb" in crypttab(5) and/or
> >> fstab(5) entries rather than the conventional "sda".
> 
> > This is news to me, and I'm not sure why you blame the d-i's going
> > against decades of precedent. Yes, it can happen, and I guess it might
> > be the result of the way the buses are configured inside the machine.
> > Gone are the simple days of PATAs on hda.
> >
> > Anyway, here are six machines: three are laptops, three EFIs, one AiO,
> > one i386, all GPT, all netinst, variously 11.1–3 with firmware. All
> > except the last have the installer on sdb, and the reason there is
> > obvious.

I'll factorize the six cases. But for the last, they were listed in
date order, and I won't requote them in full.

Jan  1 bullseye 11.1 amd64 GPT EFI  rust laptop
Jan 25 bullseye 11.2 amd64 GPT BIOS rust tower
Feb  1 bullseye 11.2 amd64 GPT BIOS rust minitower
Apr 17 bullseye 11.3 amd64 GPT EFI  rust AiO
Apr 21 bullseye 11.2 i386  GPT BIOS rust laptop
Dec 13 bullseye 11.1 amd64 GPT EFI  SSD  laptop

> > The sole occasion where the devices have been named the "wrong" way
> > round is when I installed Debian on an EFI laptop onto a virgin
> > external hard drive. If memory serves, the d-i was connected through
> > a hub to the USB-C port, the drive was on the single USB3 port.
> > I think.

[ … snipped the listings … ]

> I am unsure what is obvious about the last three.  Please clarify.

No, just the last /one/. The USB stick gets names sda because
there's no competition from the SSD, which is nvme0n1.

> Did you install from a USB flash drive or from optical media?  If
> optical, was the drive internal or external, and what was the
> interface?

They're all booted off the same USB stick:

  2006528 512-byte logical blocks: (1.03 GB/980 MiB)

which is an eight-year-old freebee courtesy of ALS Empirica
at a Houston conference. If you're wondering why it sprouted
an extra partition for Apr 21, look no further than:

https://lists.debian.org/debian-user/2022/04/msg00655.html

I wanted to confirm that a 2003-vintage laptop would BIOS boot
from a GPT disk. I now have no MBR hard-drives in the house.
But to kill two birds with one stone, I was confirming that
creating an extra partition on a d-i stick is as simple as
running fdisk and creating one in the usual way:

$ sudo fdisk -l /dev/sdb
Disk /dev/sdb: 979.75 MiB, 1027342336 bytes, 2006528 sectors
Disk model: Card  Reader    
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: dos
Disk identifier: 0x5ec18678

Device     Boot   Start     End Sectors  Size Id Type
/dev/sdb1  *          0 1155071 1155072  564M  0 Empty
/dev/sdb2          4316    8283    3968  1.9M ef EFI (FAT-12/16/32)
/dev/sdb3       1163264 2002943  839680  410M  c W95 FAT32 (LBA)
$ ls -l /dev/disk/by-label/
total 0
lrwxrwxrwx 1 root root 10 May  9 14:44 'Debian\x2011.2.0\x20i386\x20n' -> ../../sdb1
lrwxrwxrwx 1 root root 10 May  9 14:44  I386-11.2 -> ../../sdb3
lrwxrwxrwx 1 root root 10 May  9 08:40  cryptswap -> ../../sda3
lrwxrwxrwx 1 root root 10 May  9 08:40  swap -> ../../dm-0
lrwxrwxrwx 1 root root 10 May  9 08:40  toto04 -> ../../sda4
lrwxrwxrwx 1 root root 10 May  9 08:40  toto05 -> ../../sda5
lrwxrwxrwx 1 root root 10 May  9 08:40  toto06 -> ../../dm-1
$ 

The d-i runs as per usual. Whether the extra partition is usable
for, say, firmware, I haven't tested.

(ISTR creating an extra partition being discussed a while back.)

Cheers,
David.

[toc] | [prev] | [next] | [standalone]

#248059

FromDavid Christensen <dpchrist@holgerdanske.com>
Date2022-05-10 02:40 +0200
Message-ID<Elm4V-eqRO-5@gated-at.bofh.it>
In reply to#248050
On 5/9/22 12:58, David Wright wrote:
> On Mon 09 May 2022 at 11:30:39 (-0700), David Christensen wrote:
>> On 5/9/22 10:21, David Wright wrote:
>>> On Sun 08 May 2022 at 23:39:31 (-0700), David Christensen wrote:
>>
>>>> As noted by another reader in another thread, the Debian 11 installer
>>>> appears to always assign /dev/sda to d-i USB installation media; in
>>>> spite of decades of standard practice of using the first drive node
>>>> for the target drive.  Thus, the target drive may have been /dev/sdb
>>>> at installation time, resulting in "sdb" in crypttab(5) and/or
>>>> fstab(5) entries rather than the conventional "sda".
>>
>>> This is news to me, and I'm not sure why you blame the d-i's going
>>> against decades of precedent. Yes, it can happen, and I guess it might
>>> be the result of the way the buses are configured inside the machine.
>>> Gone are the simple days of PATAs on hda.
>>>
>>> Anyway, here are six machines: three are laptops, three EFIs, one AiO,
>>> one i386, all GPT, all netinst, variously 11.1–3 with firmware. All
>>> except the last have the installer on sdb, and the reason there is
>>> obvious.

>> I am unsure what is obvious about the last three.  Please clarify.
> 
> No, just the last /one/. The USB stick gets names sda because
> there's no competition from the SSD, which is nvme0n1.


Okay.  That makes sense.


>> Did you install from a USB flash drive or from optical media?  If
>> optical, was the drive internal or external, and what was the
>> interface?
> 
> They're all booted off the same USB stick ...


I just did a few trials with debian-11.3.0-amd64-netinst.iso on a 
different laptop with a different USB target drive.  The target drive 
came up as /dev/sda in each case (!).  The results do not match my 
earlier trials on other computers and targets.


I can only conclude that this is a multi-dimensional problem where the 
target computer (and firmware) is one dimension and the target drive is 
an additional dimension.


David

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.user

csiph-web